This project proposes to develop and evaluate a hand rehabilitation system for chronic hemiparetic stroke survivors that have lost their ability to full control their hand. Stroke is the leading cause of adult disability, and is the third leading cause of death in the United States. Stroke survivors often face a significant loss of functional motor control in the upper limbs of the contralesional side of the body. Restoration of upper limb function is particularly difficult due to its enormous complexity. User control of hand movement is crucial to regaining correct central motor control. It has been shown that the paralyzed hand can be improved through repetitive training of isolated movements. This program will direct development efforts towards a hand rehabilitation system that provides adaptive assistance as needed during physical rehabilitation based on feedback from both the user's volitional movement and surface electromyogram (EMG) measurements. Pattern recognition techniques will be utilized with the EMG input to determine the user's intended complex hand movement. User input is incorporated during therapeutic training maneuvers to set the appropriate level of assistance to promote efficient motor learning in stroke survivors.